itsm-email-recommendation
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes untrusted content from ServiceNow records which may contain adversarial instructions.
- Ingestion points: Data is pulled from the
sys_email(Step 2) andsys_journal_field(Step 3) tables, which store user-supplied communications and comments. - Boundary markers: There are no explicit instructions or delimiters provided to the agent to treat the retrieved content as untrusted data or to ignore embedded instructions.
- Capability inventory: The skill utilizes tools like
SN-Execute-Background-Script,SN-Update-Record, andSN-Add-Work-Notesto perform actions based on the analysis. - Sanitization: No validation or filtering is applied to the retrieved text before it is interpolated into templates or processed by the AI.
- [DYNAMIC_EXECUTION]: The skill generates and executes JavaScript code within the ServiceNow environment to aggregate record context.
- Evidence: Step 5 and Example 3 use the
SN-Execute-Background-Scripttool to run GlideRecord queries defined by string templates in the skill body. - Evidence: The skill uses variable interpolation (e.g.,
[incident_sys_id]) to customize these scripts at runtime, which is a form of dynamic execution.
Audit Metadata